CHANGELOG

=== HEAD

* Support multiple SQL statements in one query in the MySQL adapter (jeremyevans)

* Add stored procedure support for the MySQL and JDBC adapters (jeremyevans, krsgoss) (#252)

* Support options when altering a column's type (for changing enums, varchar size, etc.) (jeremyevans)

* Support AliasedExpressions in tables when using implicitly qualified arguments in joins (jeremyevans)

* Support Dataset#except on Oracle (jeremyevans)

* Raise errors when EXCEPT/INTERSECT is used when not supported (jeremyevans)

* Fix ordering of UNION, INTERSECT, and EXCEPT statements (jeremyevans) (#253)

* Support aliasing subselects in the Oracle adapter (jeremyevans)

* Add a subadapter for the Progress RDBMS to the ODBC adapter (:db_type=>'progress') (groveriffic) (#251)

* Make MySQL and Oracle adapters raise an Error if asked to do a SELECT DISTINCT ON (jeremyevans)

* Set standard_conforming_strings = ON by default when using PostgreSQL, turn off with Sequel::Postgres.force_standard_strings = false (jeremyevans) (#247)

* Fix Database#rename_table when using PostgreSQL (jeremyevans) (#248)

* Whether to upcase or quote identifiers can now be set separately, via Sequel.upcase_identifiers= or the :upcase_identifiers database option (jeremyevans)

* Support transactions in the ODBC adapter (dlee)

* Support multi_insert_sql and unicode string literals in MSSQL shared adapter (dlee)

* Make PostgreSQL use the default schema if parsing the schema for all tables at once, even if :schema=>nil option is used (jeremyevans)

* Make MySQL adapter not raise an error when giving an SQL::Identifier object to the schema modification methods such as create_table (jeremyevans)

* The keys of the hash returned by Database#schema without a table name are now quoted strings instead of symbols (jeremyevans)

* Make Database#schema to handle implicit schemas on all databases and multiple identifier object types (jeremyevans)

* Remove Sequel.odbc_mssql method (jeremyevans) (#249)

* More optimization of Model#initialize (jeremyevans)

* Treat interval as it's own type, not an integer type (jeremyevans)

* Allow use of implicitly qualified symbol as argument to Symbol#qualify (:a.qualify(:b__c)=>b.c.a), fixes model associations in different schemas (jeremyevans) (#246)

=== 2.7.1 (2008-11-04)

* Fix PostgreSQL Date optimization so that it doesn't reject dates like 11/03/2008 (jeremyevans)

=== 2.7.0 (2008-11-03)

* Transform AssociationReflection from a single class to a class hierarchy (jeremyevans)

* Optimize Date object creation in PostgreSQL adapter (jeremyevans)

* Allow easier creation of custom association types, though support for them may still be suboptimal (jeremyevans)

* Add :eager_grapher option to associations, which the user can use to override the default eager_graph code (jeremyevans)

* Associations are now inherited when a model class is subclassed (jeremyevans)

* Instance methods added by associations are now added to an anonymous module the class includes, allowing you to override them and use super (jeremyevans)

* Add #add_graph_aliases (select_more for graphs), and allow use of arbitrary expressions when graphing (jeremyevans)

* Fix a corner case where the wrong table name is used in eager_graph (jeremyevans)

* Make Dataset#join_table take an option hash instead of a table_alias argument, add support for :implicit_qualifier option (jeremyevans)

* Add :left_primary_key and :right_primary_key options to many_to_many associations (jeremyevans)

* Add :primary_key option to one_to_many and many_to_one associations (jeremyevans)

* Make after_load association callbacks take effect when eager loading via eager (jeremyevans)

* Add a :uniq association option to many_to_many associations (jeremyevans)

* Support using any expression as the argument to Symbol#like (jeremyevans)

* Much better support for multiple schemas in PostgreSQL (jeremyevans) (#243)

* The first argument to Model#initalize can no longer be nil, it must be a hash if it is given (jeremyevans)

* Remove Sequel::Model.lazy_load_schema= setting (jeremyevans)

* Lazily load model instance options such as raise_on_save_failure, for better performance (jeremyevans)

* Make Model::Validiation::Errors more Rails-compatible (jeremyevans)

* Refactor model hooks for performance (jeremyevans)

* Major performance enhancement when fetching rows using PostgreSQL (jeremyevans)

* Don't typecast serialized columns in models (jeremyevans)

* Add Array#sql_array to handle ruby arrays of all two pairs as SQL arrays (jeremyevans) (#245)

* Add ComplexExpression#== and #eql?, for checking equality (rubymage) (#244)

* Allow full text search on PostgreSQL to include rows where a search column is NULL (jeremyevans)

* PostgreSQL full text search queries with multiple columns are joined with space to prevent joining border words to one (michalbugno)

* Don't modify a dataset's cached column information if calling #each with an option that modifies the columns (jeremyevans)

* The PostgreSQL adapter will now generally default to using a unix socket in /tmp if no host is specified, instead of a tcp socket to localhost (jeremyevans)

* Make Dataset#sql call Dataset#select_sql instead of being an alias, to allow for easier subclassing (jeremyevans)

* Split Oracle adapter into shared and unshared parts, so Oracle is better supported when using JDBC (jeremyevans)

* Fix automatic loading of Oracle driver when using JDBC adapter (bburton333) (#242)

=== 2.6.0 (2008-10-11)

* Make the sqlite adapter respect the Sequel.datetime_class setting, for timestamp and datetime types (jeremyevans)

* Enhance the CASE statement support to include an optional expression (jarredholman)

* Default to using the simple language if no language is specified for a full text index on PostgreSQL (michalbugno)

* Add Model.raise_on_typecast_failure=, which makes it possible to not raise errors on invalid typecasts (michalbugno)

* Add schema.rdoc file, which provides an brief description of the various parts of Sequel related to schema modification (jeremyevans)

* Fix constraint generation when not using a proc or interpolated string (jeremyevans)

* Make eager_graph respect associations' :order options (use :order_eager_graph=>false to disable) (jeremyevans)

* Cache negative lookup when eagerly loading many_to_one associations where no objects have an associated object (jeremyevans)

* Allow string keys to be used when using Dataset#multi_insert (jeremyevans)

* Fix join_table when doing the first join for a dataset where the first source is a dataset when using unqualified columns (jeremyevans)

* Fix a few corner cases in eager_graph (jeremyevans)

* Support transactions on MSSQL (jeremyevans)

* Use string literals in AS clauses on SQLite (jeremyevans) (#241)

* AlterTableGenerator#set_column_allow_null was added to SET/DROP NOT NULL for columns (divoxx)

* Database#tables now works for MySQL databases using the JDBC adapter (jeremyevans)

* Database#drop_view can now take multiple arguments to drop multiple views at once (jeremyevans)

* Schema modification methods (e.g. drop_table, create_table!) now remove the cached schema entry (jeremyevans)

* Models can now determine their primary keys by looking at the schema (jeremyevans)

* No longer include :numeric_precision and :max_chars entries in the schema column hashes, use the :db_type entry instead (jeremyevans)

* Make schema parsing on PostgreSQL handle implicit schemas (e.g. schema(:schema__table)), so it works with models for tables outside the public schema (jeremyevans)

* Significantly speed up schema parsing on MySQL (jeremyevans)

* Include primary key information when parsing the schema (jeremyevans)

* Fix schema generation of composite foreign keys on MySQL (clivecrous, jeremyevans)

=== 2.5.0 (2008-09-03)

* Add Dataset #set_defaults and #set_overrides, used for scoping the values used in insert/update statements (jeremyevans)

* Allow Models to use the RETURNING clause when inserting records on PostgreSQL (jeremyevans)

* Raise Sequel::DatabaseError instead of generic Sequel::Error for database errors, don't swallow tracebacks (jeremyevans)

* Use INSERT ... RETURNING ... with PostgreSQL 8.2 and higher (jeremyevans)

* Make insert_sql, delete_sql, and update_sql respect the :sql option (jeremyevans)

* Default to converting 2 digit years, use Sequel.convert_two_digit_years = false to get back the old behavior (jeremyevans)

* Make the PostgreSQL adapter with the pg driver use async_exec, so it doesn't block the entire interpreter (jeremyevans)

* Make the schema generators support composite primary and foreign keys and unique constraints (jarredholman)

* Work with the 2008.08.17 version of the pg gem (erikh)

* Disallow abuse of SQL function syntax for types (use :type=>:varchar, :size=>255 instead of :type=>:varchar[255]) (jeremyevans)

* Quote index names when creating or dropping indexes (jeremyevans, SanityInAnarchy)

* Don't have column accessor methods override plugin instance methods (jeremyevans)

* Allow validation of multiple attributes at once, with built in support for uniqueness checking of multiple columns (jeremyevans)

* In PostgreSQL adapter, fix inserting a row with a primary key value inside a transaction (jeremyevans)

* Allow before_save and before_update to affect the columns saved by save_changes (jeremyevans)

* Make Dataset#single_value work when graphing, which fixes count and paginate on graphed datasets (jeremyevans)

=== 2.4.0 (2008-08-06)

* Handle Java::JavaSql::Date type in the JDBC adapter (jeremyevans)

* Add support for read-only slave/writable master databases and database sharding (jeremyevans)

* Remove InvalidExpression, InvalidFilter, InvalidJoinType, and WorkerStop exceptions (jeremyevans)

* Add prepared statement/bound variable support (jeremyevans)

* Fix anonymous column names in the ADO adapter (nusco)

* Remove odbc_mssql adapter, use :db_type=>'mssql' option instead (jeremyevans)

* Split MSSQL specific syntax into separate file, usable by ADO and ODBC adapters (nusco, jeremyevans)

=== 2.3.0 (2008-07-25)

* Enable almost full support for MySQL using JDBC (jeremyevans)

* Fix ODBC adapter's conversion of ::ODBC::Time values (Michael Xavier)

* Enable full support for SQLite-JDBC using the JDBC adapter (jeremyevans)

* Minor changes to allow for full Ruby 1.9 compatibility (jeremyevans)

* Make Database#disconnect work for the ADO adapter (spicyj)

* Don't raise an exception in the ADO adapter if the dataset contains no records (nusco)

* Enable almost full support of PostgreSQL-JDBC using the JDBC adapter (jeremyevans)

* Remove Sequel::Worker (jeremyevans)

* Make PostgreSQL adapter not raise an error when inserting records into a table without a primary key (jeremyevans)

* Make Database.uri_to_options a private class method (jeremyevans)

* Make JDBC load drivers automatically for PostgreSQL, MySQL, SQLite, Oracle, and MSSQL (jeremyevans)

* Make Oracle adapter work with a nonstandard Oracle database port (pavel.lukin)

* Typecast '' to nil by default for non-string non-blob columns, add typecast_empty_string_to_nil= model class and instance methods (jeremyevans)

* Use a simpler select in Dataset#empty?, fixes use with MySQL (jeremyevans)

* Add integration test suite, testing sequel against a real database, with nothing mocked (jeremyevans)

* Make validates_length_of default tag depend on presence of options passed to it (jeremyevans)

* Combine the directory structure for sequel_model and sequel_core, now there is going to be only one gem named sequel (jeremyevans)

=== 2.2.0 (2008-07-05)

* Add :extend association option, extending the dataset with module(s) (jeremyevans)

* Add :after_load association callback option, called after associated objects have been loaded from the database (jeremyevans)

* Make validation methods support a :tag option, to work correctly with source reloading (jeremyevans)

* Add :before_add, :after_add, :before_remove, :after_remove association callback options (jeremyevans)

* Break many_to_one association setter method in two parts, for easier overriding (jeremyevans)

* Model.validates_presence_of now considers false as present instead of absent (jeremyevans)

* Add Model.raise_on_save_failure, raising errors on save failure instead of return false (now nil), default to true (jeremyevans)

* Add :eager_loader association option, to specify code to be run when eager loading (jeremyevans)

* Make :many_to_one associations support :dataset, :order, :limit association options, as well as block arguments (jeremyevans)

* Add :dataset association option, which overrides the default base dataset to use (jeremyevans)

* Add :eager_graph association option, works just like :eager except it uses #eager_graph (jeremyevans)

* Add :graph_join_table_join_type association option (jeremyevans)

* Add :graph_only_conditions and :graph_join_table_only_conditions association options (jeremyevans)

* Add :graph_block and :graph_join_table_block association options (jeremyevans)

* Set the model's dataset's columns in addition to the model's columns when loading the schema for a model (jeremyevans)

* Make caching work correctly with subclasses (jeremyevans)

* Add the Model.to_hash dataset method (jeremyevans)

* Filter blocks now yield a SQL::VirtualRow argument, which is useful if another library defines operator methods on Symbol (jeremyevans)

* Add Symbol#identifier method, to make x__a be treated as "x__a" instead of "x"."a" (jeremyevans)

* Dataset#update no longer takes a block, please use a hash argument with the expression syntax instead (jeremyevans)

* ParseTree support has been removed from Sequel (jeremyevans)

* Database#drop_column is now supported in the SQLite adapter (abhay)

* Tinyint columns can now be considered integers instead of booleans by setting Sequel.convert_tinyint_to_bool = false (samsouder)

* Allow the use of URL parameters in connection strings (jeremyevans)

* Ignore any previously selected columns when using Dataset#graph for the first time (jeremyevans)

* Dataset#graph now accepts a block which is passed to join_table (jeremyevans)

* Make Dataset#columns ignore any filtering, ordering, and distinct clauses (jeremyevans)

* Use the safer connection-specific string escaping methods for PostgreSQL (jeremyevans)

* Database#transaction now yields a connection when using the Postgres adapter, just like it does for other adapters (jeremyevans)

* Dataset#count now works for a limited dataset (divoxx)

* Database#add_index is now supported in the SQLite adapter (abhay)

* Sequel's MySQL adapter should no longer conflict with ActiveRecord's use of MySQL (careo)

* Treat Hash as expression instead of column alias when used in DISTINCT, ORDER BY, and GROUP BY clauses (jeremyevans)

* PostgreSQL bytea fields are now fully supported (dlee)

* For PostgreSQL, don't raise an error when assigning a value to a SERIAL PRIMARY KEY field when inserting records (jeremyevans)

=== 2.1.0 (2008-06-17)

* Break association add_/remove_/remove_all_ methods into two parts, for easier overriding (jeremyevans)

* Add Model.strict_param_setting, on by default, which raises errors if a missing/restricted method is called via new/set/update/etc. (jeremyevans)

* Raise errors when using association methods on objects without valid primary keys (jeremyevans)

* The model's primary key is a restricted column by default, Add model.unrestrict_primary_key to get the old behavior (jeremyevans)

* Add Model.set_(allowed|restricted)_columns, which affect which columns create/new/set/update/etc. modify (jeremyevans)

* Calls to Model.def_dataset_method with a block are cached and reapplied to the new dataset if set_dataset is called, even in a subclass (jeremyevans)

* The :reciprocal option to associations should now be the symbol name of the reciprocal association, not an instance variable symbol (jeremyevans)

* Add Model#associations, which is a hash holding a cache of associated objects, with each association being a separate key (jeremyevans)
  
* Make all associations support a :graph_select option, specifying a column or array of columns to select when using eager_graph (jeremyevans)

* Bring back Model#set and Model#update, now the same as Model#set_with_params and Model#update_with_params (jeremyevans)

* Allow model datasets to call to_hash without any arguments, which allows easy creation of identity maps (jeremyevans)

* Add Model.set_sti_key, for easily setting up single table inheritance (jeremyevans)

* Make all associations support a :read_only option, which doesn't add methods that modify the database (jeremyevans)

* Make *_to_many associations support a :limit option, for specifying a limit to the resulting records (and possibly an offset) (jeremyevans)

* Make association block argument and :eager option affect the _dataset method (jeremyevans)

* Add a :one_to_one option to one_to_many associations, which creates a getter and setter similar to many_to_one (a.k.a. has_one) (jeremyevans)

* add_ and remove_ one_to_many association methods now raise an error if the passed object cannot be saved, instead of saving without validation (jeremyevans)

* Add support for :if option on validations, using a symbol (specifying an instance method) or a proc (dtsato)

* Support bitwise operators for NumericExpressions: &, |, ^, ~, <<, >> (jeremyevans)

* No longer raise an error for Dataset#filter(true) or Dataset#filter(false) (jeremyevans)

* Allow Dataset #filter, #or, #exclude and other methods that call them to use both the block and regular arguments (jeremyevans)

* ParseTree support is now officially deprecated, use Sequel.use_parse_tree = false to use the expression (blockless) filters inside blocks (jeremyevans)

* Remove :pool_reuse_connections ConnectionPool/Database option, MySQL users need to be careful with nested queries (jeremyevans)

* Allow Dataset#graph :select option to take an array of columns to select (jeremyevans)

* Allow Dataset#to_hash to be called with only one argument, allowing for easy creation of lookup tables for a single key (jeremyevans)

* Allow join_table to accept a block providing the aliases and previous joins, that allows you to specify arbitrary conditions properly qualified (jeremyevans)

* Support NATURAL, CROSS, and USING joins in join_table (jeremyevans)

* Make sure HAVING comes before ORDER BY, per the SQL standard and at least MySQL, PostgreSQL, and SQLite (juco)

* Add cast_numeric and cast_string methods for use in the Sequel DSL, that have default types and wrap the object in the correct class (jeremyevans)

* Add Symbol#qualify, for adding a table/schema qualifier to a column/table name (jeremyevans)

* Remove Module#metaprivate, since it duplicates the standard Module#private_class_method (jeremyevans)

* Support the SQL CASE expression via Array#case and Hash#case (jeremyevans)

* Support the SQL EXTRACT function: :date.extract(:year) (jeremyevans)

* Convert numeric fields to BigDecimal in PostgreSQL adapter (jeremyevans)

* Add :decimal fields to the schema parser (jeremyevans)

* The expr argument in join table now allows the same argument as filter, so it can take a string or a blockless filter expression (brushbox, jeremyevans)

* No longer assume the expr argument to join_table references the primary key column (jeremyevans)

* Rename the Sequel.time_class setting to Sequel.datetime_class (jeremyevans)

* Add savepoint/nesting support to postgresql transactions (elven)

* Use the specified table alias when joining a dataset, instead of the automatically generated alias (brushbox)

=== 2.0.1 (2008-06-04)

* Make the choice of Time or DateTime optional for typecasting :datetime types, default to Time (jeremyevans)

* Reload database schema for table when calling Model.create_table (jeremyevans)

* Have PostgreSQL money type use BigDecimal instead of Float (jeremyevans)

* Have the PostgreSQL and MySQL adapters use the Sequel.time_class setting for datetime/timestamp types (jeremyevans)

* Add Sequel.time_class and String#to_sequel_time, used for converting time values from the database to either Time (default) or DateTime (jeremyevans)

* Make identifier quoting uppercase by default, to work better with the SQL standard, override in PostgreSQL (jeremyevans) (#232)

* Add StringExpression#+, for simple SQL string concatenation (:x.sql_string + :y) (jeremyevans)

* Make StringMethods.like to a case sensensitive search on MySQL (use ilike for the old behavior) (jeremyevans)

* Add StringMethods.ilike, for case insensitive pattern matching (jeremyevans)

* Refactor ComplexExpression into three subclasses and a few modules, so operators that don't make sense are not defined for the class (jeremyevans)

=== 2.0.0 (2008-06-01)

* Comprehensive update of all documentation (jeremyevans)

* Remove methods deprecated in 1.5.0 (jeremyevans)

* Add typecasting on attribute assignment to Sequel::Model objects, optional but enabled by default (jeremyevans)

* Returning false in one of the before_ hooks now causes the appropriate method(s) to immediately return false (jeremyevans)

* Add remove_all_* association method for *_to_many associations, which removes the association with all currently associated objects (jeremyevans)

* Add Model.lazy_load_schema=, when set to true, it loads the schema on first instantiation (jeremyevans)

* Add before_validation and after_validation hooks, called whenever the model is validated (jeremyevans)

* Add Model.default_foreign_key, a private class method that allows changing the default foreign key that Sequel will use in associations (jeremyevans)

* Cache negative lookup when eagerly loading many_to_one associations (jeremyevans)

* Make all associations support the :select option, not just many_to_many (jeremyevans)

* Allow the use of blocks when eager loading, and add the :eager_block and :allow_eager association options for configuration (jeremyevans)
  
* Add the :graph_join_type, :graph_conditions, and :graph_join_table_conditions association options, used when eager graphing (jeremyevans)

* Add AssociationReflection class (subclass of Hash), to make calling a couple of private Model methods unnecessary (jeremyevans)

* Change hook methods so that if a tag/method is specified it overwrites an existing hook block with the same tag/method (jeremyevans)

* Refactor String inflection support, you must use String.inflections instead of Inflector.inflections now (jeremyevans)

* Allow connection to ODBC-MSSQL via a URL (petersumskas) (#230)

* Comprehensive update of all documentation, except for the block filters and adapters (jeremyevans)

* Handle Date and DateTime value literalization correctly in adapters (jeremyevans)

* Literalize DateTime values the same as Time values (jeremyevans)

* MySQL tinyints are now returned as boolean values instead of integers (jeremyevans)

* Set additional MySQL charset options required for creating tables and databases (tmm1)

* Remove methods deprecated in 1.5.0 (jeremyevans)

* Add Module#metaattr_accessor for creating attr_accessors for the metaclass (jeremyevans)

* Add SQL string concatenation support to blockless filters, via Array#sql_string_join (jeremyevans)

* Add Pagination#last_page? and Pagination#first_page? (apeiros)

* Add limited column reflection support, tested on PostgreSQL, MySQL, and SQLite (jeremyevans)

* Allow the use of :schema__table___table_alias syntax for tables, similar to the column support (jeremyevans)

* Merge metaid gem into core_ext.rb and clean it up, so sequel now has no external dependencies (jeremyevans)

* Add Dataset#as, so using a dataset as a column with an alias is not deprecated (jeremyevans)

* Add Dataset#invert, which returns a dataset with inverted HAVING and WHERE clauses (jeremyevans)

* Add blockless filter syntax support (jeremyevans)

* Passing an array to Dataset#order and Dataset#select no longer works, you need to pass multiple arguments (jeremyevans)

* You should use '?' instead of '(?)' when using interpolated strings with array arguments (jeremyevans)

* Dataset.literal now surrounds the literalization of arrays with parentheses (jeremyevans)

* Add echo option (back?) to sequel command line tool, via -E or --echo (jeremyevans)

* Allow databases to have multiple loggers (jeremyevans)

* The sequel command line tool now also accepts a path to a database config YAML file in addition to a URI (mtodd)

* Major update of the postgresql adapter (jdavis, jeremyevans) (#225)

* Make returning inside of a database transaction commit the transaction (ahoward, jeremyevans)

* Dataset#to_table_reference is now protected, and it has a different API (jeremyevans)

* Dataset#join_table and related functions now take an explicit optional table_alias argument, you can no longer include the table alias in the table argument (jeremyevans)

* Aliased and/or qualified columns with embedded spaces can now be specified as symbols (jeremyevans)

* When identifier quoting is enabled, the SQL standard double quote is used by default (jeremyevans)

* When identifier quoting is enabled, quote tables as well as columns (jeremyevans)

* Make identifier quoting optional, enabled by default (jeremyevans)

* Allow Sequel::Database.connect and related methods to take a block that disconnects the database when the block finishes (jeremyevans)

* Add Dataset#unfiltered, for removing filters from dataset (jeremyevans)

* Add add_foreign_key and add_primary_key methods to the AlterTableGenerator (jeremyevans)

* Allow migration files to have more than 3 digits (jeremyevans)

* Add methods directly to Dataset instead of including modules (jeremyevans)

* Make some Dataset instance methods private: invert_order, insert_default_values_sql (jeremyevans)

* Don't add methods that depend on ParseTree unless you can load ParseTree (jeremyevans)

* Don't wipeout the cached columns every time a dataset is cloned, but only on changes to :select, :sql, :from, or :join (jeremyevans)

* Fix Oracle Adapter (yasushi.abe)

* Fixed sqlite uri so that sqlite:// works just like file:// (2 slashes for a relative path, 3 for an absolute) (dlee)

* Raise a Sequel::Error if an invalid limit or offset is used (jeremyevans)

* Refactor and beef up Dataset#first and Dataset#last, with some change in functionality (jeremyevans)

* Add String#to_datetime, for consistency (jeremyevans)

* Fix Range#interval so that it returns 1 less for an exclusive range

* Change SQLite adapter so it doesn't swallow exceptions other than SQLite3::Exception (such as Interrupt) (jeremyevans)

* Change PostgreSQL and MySQL adapters to raise Sequel::Error instead of database specific errors if a database error occurs (jeremyevans)

* Using a memory database with SQLite now defaults to a single connection, so all queries it uses run against the same database (jeremyevans)

* Fix attempting to query MySQL using the same connection being used to concurrently execute another query (jeremyevans)

* Add options to the connection pool to configure reusing connections and converting exceptions (jeremyevans)

* Use the database driver provided string quoting methods for MySQL and SQLite (jeremyevans) (#223)

* Add ColumnAll#==, for checking the equality of two ColumnAlls (jeremyevans)

* Allow an array of arrays instead of a hash when specifying conditions (jeremyevans)

* Add Sequel::DBI::Database#lowercase, for lowercasing column names (jamesearl)

* Remove Dataset#extend_with_destroy, which may break code that uses Dataset#set_model directly and expects the destroy method to be added (jeremyevans)

* Fix some issues when running on Ruby 1.9 (Zverok, jeremyevans)

* Make the DBI adapter work (partially) with PostgreSQL (Seb)

=== 1.5.1 (2008-04-30)

* Fix Dataset#eager_graph when not all objects have associated objects (jeremyevans)

* Have Dataset#graph give a nil value instead of a hash with all nil values if no matching rows exist in the graphed table (jeremyevans)

=== 1.5.0 (2008-04-29)

* Make the validation errors API compatible with Merb (Inviz)

* Add validates_uniqueness_of, for protecting against duplicate entries in the database (neaf, jeremyevans)

* Alias Model#dataset= to Model#set_dataset (tmm1)

* Make some Model class methods private: def_hook_method, hooks, add_hook, plugin_module, plugin_gem (jeremyevans)

* Add the eager! and eager_graph! mutation methods to model datasets (jeremyevans)

* Remove Model.database_opened (jeremyevans)

* Remove Model.super_dataset (jeremyevans)

* Deprecate .create_with_params, .create_with, #set, #update, #update_with, and #new_record from Sequel::Model (jeremyevans)

* Add Model.def_dataset_method, for defining methods on the model that reference methods on the dataset (jeremyevans)

* Deprecate Model.method_missing, add dataset methods to Model via metaprogramming (jeremyevans)

* Remove Model.join, so it is the same as Dataset#join (jeremyevans)

* Use reciprocal associations for all types of associations in the getter/setter/add_/remove_ methods (jeremyevans)

* Fix many_to_one associations to cache negative lookups (jeremyevans)

* Change Model#=== to always be false if the primary key is nil (jeremyevans)

* Add Model#hash, which should be unique for a given class and primary key (or values if primary key is nil) (jeremyevans)

* Add Model#eql? as a alias to Model#== (jeremyevans)

* Make Model#reload clear any cached associations (jeremyevans)

* No longer depend on the assistance gem, merge the Inflector and Validations code (jeremyevans)

* Add Model#set_with_params, which is Model#update_with_params without the save (jeremyevans)
  
* Fix Model#destroy so that it returns self, not the result of after_destroy (jeremyevans)

* Define Model column accessors in set_dataset, so they should always be avaiable, deprecate Model#method_missing (jeremyevans)

* Add eager loading of associations via new sequel_core object graphing feature (jeremyevans)

* Fix many_to_many associations with classes inside modules without an explicit join table (jeremyevans)

* Allow creation of new records that don't have primary keys when the cache is on (jeremyevans) (#213)

* Make Model#initialize, Model#set, and Model#update_with_params invulnerable to memory exhaustion (jeremyevans) (#210)

* Add Model.str_columns, which gives a list of columns as frozen strings (jeremyevans)

* Remove pretty_table.rb from sequel, since it is in sequel_core (jeremyevans)

* Set a timeout in the Sqlite adapter, default to 5 seconds (hrvoje.marjanovic) (#218)

* Document that calling Sequel::ODBC::Database#execute manually requires you to manually drop the returned object (jeremyevans) (#217)

* Paginating an already paginated/limited dataset now raises an error (jeremyevans)

* Add support for PostgreSQL partial indexes (dlee)

* Added support for arbitrary index types (including spatial indexes) (dlee)

* Quote column names in SQL generated for SQLite (tmm1)

* Deprecate Object#rollback! (jeremyevans)

* Make some Dataset methods private (qualified_column_name, column_list, table_ref, source_list) (jeremyevans)

* Deprecate Dataset methods #set_options, #set_row_proc, #remove_row_proc, and #clone_merge (jeremyevans)

* Add Symbol#*, a replacement for Symbol#all (jeremyevans)

* Deprecate including ColumnMethods in Object, include it in Symbol, String, and Sequel::SQL::Expression (jeremyevans)

* Deprecate Symbol#method_missing, and #AS, #DESC, #ASC, #ALL, and #all from ColumnMethods (jeremyevans)

* Fix table joining in MySQL (jeremyevans)

* Deprecate Sequel.method_missing and Object#Sequel, add real Sequel.adapter methods (jeremyevans)

* Move dataset methods applicable only to paginated datasets into Sequel::Dataset::Pagination (jeremyevans)

* Make Sequel::Dataset::Sequelizer methods private (jeremyevans)

* Deprecate Dataset#method_missing, add real mutation methods (e.g. filter!) (jeremyevans)

* Fix connecting to an MSSQL server via ODBC using domain user credentials (jeremyevans) (#216)

* No longer depend on the assistance gem, merge in the ConnectionPool and .blank methods (jeremyevans)

* No longer depend on ParseTree, RubyInline, or ruby2ruby, but you still need them if you want to use the block filters (jeremyevans)

* Fix JDBC adapter by issuing index things start at 1 (pdamer)

* Fix connecting to a database via the ADO adapter (now requires options instead of URI) (timuckun, jeremyevans) (#204)

* Support storing microseconds in postgres timestamp fields (schnarch...@rootimage.msu.edu) (#215)

* Allow joining of multiple datasets, by making the table alias different for each dataset joined (jeremyevans)

* SECURITY: Fix backslash escaping of strings (dlee)

* Add ability to create a graph of objects from a query, with the result split into corresponding tables (jeremyevans) (#113)

* Add attr_accessor for dataset row_proc (jeremyevans)

* Don't redefine Dataset#each when adding a transform or row_proc (jeremyevans)

* Remove array_keys.rb from sequel_core, it was partially broken (since the arrays came from hashes), and redefined Dataset#each (jeremyevans)

* Fix MySQL default values insert (matt.binary) (#196)

* Fix ODBC adapter improperly escaping date and timestamp values (leo.borisenko) (#165)

* Fix renaming columns on MySQL with type :varchar (jeremyevans) (#206)

* Add Sequel::SQL::Function#==, for comparing SQL Functions (jeremyevans) (#209)

* Update Informix adapter to work with Ruby/Informix 0.7.0 (gerardo.santana@gmail.com)

* Remove sequel_core's knowledge of Sequel::Model (jeremyevans)

* Use "\n" instead of $/ (since $/ can be redefined in ways we do not want) (jeremyevans)

=== 1.4.0 (2008-04-08) 

* Don't mark a column as changed unless the new value is different from the current value (tamas.denes, jeremyevans) (#203).

* Switch gem name from "sequel_model" to just "sequel", which required large version bump (jeremyevans).

* Add :select option to many_to_many associations, default to selecting only the associated model table and not the join table (jeremyevans) (#208).

* Add :reciprocal one_to_many association option, for setting corresponding many_to_one instance variable (jeremyevans).

* Add eager loading implementation (jeremyevans).

* Change *_to_many associations so that the all associations are considered :cache=>true (jeremyevans).

* Fix associations with block arguments and :cache=>true (jeremyevans).

* Merge 3 mysql patches from the bugtracker (mvyver) (#200, #201, #202).

* Merge 2 postgresql patches from the bugtracker (a...@mellowtone.co.jp) (#211, 212).

* Allow overriding of default posgres spec database via ENV['SEQUEL_PG_SPEC_DB'] (jeremyevans).

* Allow using the Sequel::Model as the first argument in a dataset join selection (jeremyevans) (#170).

* Add simple callback mechanism to make model eager loading implementation easier (jeremyevans).

* Added Sequel::Error::InvalidOperation class for invalid operations (#198).

* Implemented MySQL::Database#server_version (#199).

* Added spec configuration for MySQL socket file.

* Fixed transform with array tuples in postgres adapter.

* Changed spec configuration to Database objects instead of URIs in order to support custom options for spec databases.

* Renamed schema files.

* Fixed Dataset#from to work correctly with SQL functions (#193).

===Previous to 1.4.0, Sequel model and Sequel core versioning differed, see the bottom of this file for the changelog to Sequel model prior to 1.4.0.

=== 1.3 (2008-03-08)

* Added configuration file for running specs (#186).

* Changed Database#drop_index to accept fixed arity (#173).

* Changed column definition sql to put UNSIGNED constraint before unique in order to satisfy MySQL (#171).

* Enhanced MySQL adapter to support load data local infile_, added compress option for mysql connection by default (#172).

* Fixed bug when inserting hashes in array tuples mode.

* Changed SQLite adapter to catch RuntimeError raised when executing a statement and raise an Error::InvalidStatement with the offending SQL and error message (#188).

* Added Error::InvalidStatement class.

* Fixed Dataset#reverse to not raise for unordered dataset (#189).

* Added Dataset#unordered method and changed #order to remove order if nil is specified (#190).

* Fixed reversing order of ASC expression (#164).

* Added support for :null => true option when defining table columns (#192).

* Fixed Symbol#method_missing to accept variable arity (#185).

=== 1.2.1 (2008-02-29)

* Added add_constraint and drop_constraint functionality to Database#alter_table (#182).

* Enhanced Dataset#multi_insert to accept datasets (#179).

* Added MySQL::Database#use method for switching database (#180).

* Enhanced Database.uri_to_options to accept uri strings (#178).

* Added Dataset#columns! method that always makes a roundtrip to the DB (#177).

* Added new Dataset#each_page method that iterates over all pages in the result set (#175).

* Added Dataset#reverse alias to Dataset#reverse_order (#174).

* Fixed Dataset#transform_load and #transform_save to create a trasnformed copy of the supplied hash instead of transforming it in place (#184).

* Implemented MySQL::Dataset#replace (#163).

=== 1.2 (2008-02-15)

* Added support for :varchar[100] like type declarations in #create_table.

* Fixed #rename_column in mysql adapter to support types like varchar(255) (#159).

* Added support for order and limit in DELETE statement in MySQL adapter (#160).

* Added checks to Dataset#multi_insert to prevent work if no values are given (#162).

* Override ruby2ruby implementation of Proc#to_sexp which leaks memory (#161).

* Added log option, help for sequel script (#157).

=== 1.1 (2008-02-15)

* Fixed Dataset#join_table to support joining of datasets (#156).

* Changed Dataset#empty? to use EXISTS condition instead of counting records, for much better performance (#158).

* Implemented insertion of multiple records in a single statement for postgres adapter. This feature is available only in postgres 8.2 and newer.

* Implemented Postgres::Database#server_version.

* Implemented Database#get, short for dataset.get(...).

* Refactored Dataset#multi_insert, added #import alias, added support for calling #multi_insert using array of columns and array of value arrays (thanks David Lee).

* Implemented Dataset#get, a replacement for select(column).first[column].

* Implemented Dataset#grep method, poor man's text search.

=== 1.0.10 (2008-02-13)

* Fixed Datset#group_and_count to work inside a query block (#152).

* Changed datasets with transforms to automatically transform hash filters (#155).

* Changed Marshal stock transform to use Base64 encoding with backward-compatibility to support existing marshaled values (#154).

* Added support for inserting multiple records in a single statement using #multi_insert in MySQL adapter (#153).

* Added support for :slice option (same as :commit_every) in Dataset#multi_insert.

* Changed Dataset#all to accept opts and iteration block.

=== 1.0.9 (2008-02-10)

* Implemented Dataset#inspect and Database#inspect (#151).

* Added full-text searching for odbc_mssql adapter (thanks Joseph Love).

* Added AlterTableGenerator#add_full_text_index method.

* Implemented full_text indexing and searching for PostgreSQL adapter (thanks David Lee).

* Implemented full_text indexing and searching for MySQL adapter (thanks David Lee).

* Fixed Dataset#insert_sql to work with array subscript references (thanks Jim Morris).

=== 1.0.8 (2008-02-08)

* Added support for multiple choices in string matching expressions (#147).

* Renamed Dataset#clone_merge to Dataset#clone, works with or without options for merging (#148).

* Fixed MySQL::Database#<< method to always free the result in order to allow multiple calls in a row (#149). Same also for PostgreSQL adapter.

=== 1.0.7 (2008-02-05)

* Added support for conditional filters (using if else statements) inside block filters (thanks Kee).

=== 1.0.6 (2008-02-05)

* Removed code pollution introduced in revs 814, 817 (really bad patch, IMO).

* Fixed joining datasets using aliased tables (#140).

* Added support additional field types in postgresql adapter (#146).

* Added support for date field types in postgresql adapter (#145).

* Fixed Dataset#count to work correctly for grouped datasets (#144).

* Added Dataset#select_more, Dataset#order_more methods (#129).

=== 1.0.5 (2008-01-25)

* Added support for instantiating models by using the load constructor method.

=== 1.0.4.1 (2008-01-24)

* Fixed bin/sequel to require sequel_model if available.

=== 1.0.4 (2008-01-24)

* Added Dataset#select_all method.

* Changed ODBC::Database to support connection using driver and database name, also added support for untitled columns in ODBC::Dataset (thanks Leonid Borisenko).

* Fixed MySQL adapter to correctly format foreign key definitions (#123).

* Changed MySQL::Dataset to allow HAVING clause on ungrouped datasets, and put HAVING clause before ORDER BY clause (#133).

* Changed Dataset#group_and_count to accept multiple columns (#134).

* Fixed database spec to open YAML file in binary mode (#131).

* Cleaned up gem spec (#132).

* Added Dataset#table_exists? convenience method.

=== 1.0.3 (2008-01-17)

* Added support for UNSIGNED constraint, used in MySQL? (#127).

* Implemented constraint definitions inside Database#create_table.

* Fixed postgres adapter to define PGconn#async_exec as alias to #exec if not defined (for pure-ruby postgres driver).

* Added String#to_date. Updated mysql adapter to use String#to_date for mysql date types (thanks drfreeze).

=== 1.0.2 (2008-01-14)

* Removed ConnectionPool, NumericExtensions. Added dependency on assistance.

=== 1.0.1 (2008-01-12)

* Changed postgres adapter to quote column references using double quotes.

* Applied patch for oracle adapter: fix behavior of limit and offset, transactions, #table_exists?, #tables and additional specs (thanks Liming Lian #122).

* Allow for additional filters on a grouped dataset (#119 and #120)

* Changed mysql adapter to default to localhost if :host option is not specified (#114).

* Refactored Sequelizer to use Proc#to_sexp (method provided by r2r).

* Enhanced Database.connect to accept options with string keys, so it can now accept options loaded from YAML files. Database.connect also automatically converts :username option into :user for compatibility with existing YAML configuration files for AR and DataMapper.

=== 1.0.0.1 (2008-01-03)

* Changed MySQL adapter to support specifying socket option.

* Added support for limiting and paginating datasets with fixed SQL, gotten with DB#fetch (thanks Ruy Diaz). 

* Added new Dataset#from_self method that returns a dataset selecting from the original dataset.

=== 1.0 (2008-01-02)

* Removed deprecated adapter stubs.

* Removed Sequel::Model() stub.

* Changed name to sequel_core.

* 100% code coverage.

* Fixed error behavior when sequel_model is not available.

* Fixed error behavior when parse_tree or ruby2ruby are not available.

=== 0.5.0.2 (2008-01-01)

* Fixed String#to_time to raise error correctly for invalid time stamps.

* Improved code coverage - now at 99.2%.

=== 0.5.0.1 (2007-12-31)

* Added a stub for Sequel::Model that auto-loads sequel_model.

* Changed Sequel.method_missing and Database.adapter_class to raise AdapterNotFound if an adapter could not be loaded.

* Fixed behavior of error trap in sequel command line tool.

=== 0.5 (2007-12-30)

* Removed model code into separate sub-project. Rearranged trunk into core, model and model_plugins.

=== 0.4.5 (2007-12-25)

* Added rdoc for new alter_table functionality (#109).

* Fixed update_sql with array sub-item keys (#110).

* Refactored model specs.

* Added Model#update as alias to #set.

* Refactored validations code. Renamed Model.validations? into Model.has_validations?.

* Added initial Model validations (Thanks Lance Carlson)

* Added Database#set_column_default method (thanks Jim Morris.)

* Removed warning on uninitialized @transform value (thanks Jim Morris).

=== 0.4.4.2 (2007-12-20)

* Fixed parsing errors in Ruby 1.9.

* Fixed sync problem in connection_pool_spec.

* Changed String#to_time to raise Error::InvalidValue if Time.parse fails.

* Refactored sequel error classes.

=== 0.4.4.1 (2007-12-19)

* Fixed schema generation code to use field quoting and support adapter-specific literalization of default values (#108).

=== 0.4.4 (2007-12-17)

* Implemented Database#rename_table (#104).

* Fixed drop_index in mysql adapter (#103).

* Added ALTER TABLE specs for postgres, sqlite and mysql adapters. Added custom alter_table behavior for sqlite and mysql adapters (#101, #102).

* Added direct Database API for altering tables.

* Added Database#alter_table method with support for adding, dropping, renaming, modifying columns and adding and droppping indexes.

* Added #unique schema method for defining unique indexes (thanks Dado).

* Implemented unfolding of #each calls inside sequelizer blocks (thanks Jim Morris).

=== 0.4.3 (2007-12-15)

* Fixed Dataset#update to accept strings (#98).

* Fixed Model.[] to raise for boolean argument (#97).

* Added Database#add_index method (thanks coda.hale).

* Added error reporting for filtering on comparison not in a block (thanks Jim Morris).

* Added support for inline index definition (thanks Dado).

* Added Database#create_table! method for forcibly creating a table (thanks Dado).

* Added support for using Dataset#update with block.

* Changed subscript access to use | operator.

* Fixed subscript access in sequelizer.

* Added support for subscript access using Symbol#/ operator.

=== 0.4.2.2 (2007-12-10)

* Improved code coverage.

* Fixed Dataset#count to work properly with datasets with fixed SQL (when using #fetch).

* Added Model.create_with_params method that filters the given parameters accordring to the model's columns (thanks Aman Gupta). 

=== 0.4.2.1 (2007-12-09)

* Refactored and fixed Dataset#reverse_order to work with field quoting (thanks Christian).

* Fixed problem with field quoting in insert statements.

* Changed sequelizer code to silently fail on any error when requiring parsetree and ruby2ruby.

* Added Database#create_view, #create_or_replace_view and #drop_view methods. Also implemented Dataset#create_view and #create_or_replace_view convenience methods.

* Keep DRY by re-using Model#[]= from method_missing.

* Added Model.fetch alias for DB.fetch.set_model(Model)

=== 0.4.2 (2007-12-07)

* Implemented Model#save_changes.

* Extended Model#save to accept specific columns to update.

* Implemented experimental JDBC adapter.

* Added adapter skeleton as starting point for new adapters.

* Cleaned-up adapters and moved automatic requiring of 'sequel' to adapter stubs.

=== 0.4.1.3 (2007-12-05)

* Better plugin conventions.

* Added experimental OpenBase adapter.

* Fixed Sequel.<xxx> methods to accept options hash as well as database name. Fixed Sequel.connect to accept options hash as well as URI (Wayne).

=== 0.4.1.2 (2007-12-04)

* Added release rake task (using RubyForge).

* Changed Model.is to accept variable arity.

* Implemented plugin loading for model classes.

* Fixed odbc-mssql and odbc adapters (thanks Dusty.)

* Implemented odbc-mssql adapter (thanks Dusty.)

=== 0.4.1.1 (2007-11-27)

* Fixed #first and #last functionality in Informix::Dataset (thanks Gerardo Santana).

=== 0.4.1 (2007-11-25)

* Put adapter files in lib/sequel/adapters. Requiring sequel/<adapter> is now deprecated. Users can now just require 'sequel' and adapters are automagically loaded (#93).

=== 0.4.0 (2007-11-24)

* Reorganized lib directory structure.

* Added support for dbi-xxx URI schemes (#86).

* Fixed problem in Database#uri where setting the password would raise an error (#87).

* Improved Dataset#insert_sql to correctly handle string keys (#92).

* Improved error-handling for worker threads. Errors are saved to an array and are accessible through #errors (#91).

* Dataset#uniq/distinct can now accept a column list for DISTINCT ON clauses.

* Fixed Model.all.

* Fixed literalization of strings with escape sequences in postgres adapter (#90).

* Added support for literalizing BigDecimal values (#89).

* Fixed column qualification for joined datasets (thanks Christian).

* Implemented experimental informix adapter.

=== 0.3.4.1 (2007-11-10)

* Changed Dataset#select_sql to support queries without a FROM clause.

=== 0.3.4 (2007-11-10)

* Fixed MySQL adapter to allow calling stored procedures (thanks Sebastian).

* Changed Dataset#each to always return self.

* Fixed SQL functions without arguments in block filters.

* Implemented super-cool Symbol#cast_as method.

* Fixed error message in command-line tool if failed to load adapter (#85).

* Refactored code relating to column references for better extendibility (#88).

* Tiny fix to Model#run_hooks.

=== 0.3.3 (2007-11-04)

* Revised code to generate SQL statements without trailing semicolons.

* Added Sequel::Worker implementation of a simple worker thread for asynchronous execution.

* Added spec for Oracle adapter.

* Fixed Oracle adapter to format INSERT statements without semicolons (thanks Liming Lian).

* Renamed alias to Array#keys as Array#columns instead of Array#fields.

* Renamed FieldCompositionMethods as ColumnCompositionMethods.

* Implemented Sequel::NumericExtensions to provide stuff like 30.days.ago.

=== 0.3.2 (2007-11-01)

* Added #to_column_name as alias to #to_field_name, #column_title as alias to #field_title.

* Added Dataset#interval method for getting interval between minimum/maximum values for a column.

* Fixed Oracle::Database#execute (#84).

* Added group_and_count as general implementation for count_by_xxx.

* Added count_by magic method.

* Added Dataset#range method for getting the minimum/maximum values for a column.

* Fixed timestamp translation in SQLite adapter (#83).

* Experimental DB2 adapter.

* Added Dataset#set as alias to Dataset#update.

* Removed long deprecated expressions.rb code.

* Better documentation.

* Implemented Dataset magic methods: order_by_xxx, group_by_xxx, filter_by_xxx, all_by_xxx, first_by_xxx, last_by_xxx.

* Changed Model.create and Model.new to accept a block.

=== 0.3.1 (2007-10-30)

* Typo fixes (#79).

* Added require 'yaml' to dataset.rb (#78).

* Changed postgres adapter to use the ruby-postgres library's type conversion if available (#76).

* Fixed string literalization in mysql adapter for strings with comment backslashes in them (#75).

* Fixed ParseTree dependency to work with version 2.0.0 and later (#74).

* foreign_key definitions now accept :key option for specifying the remote key (#73).

* Fixed Model#method_missing to not raise error for columns not in the table but for which a value exists (#77).

* New documentation for Model.

* Implemented Oracle adapter based on ruby-oci8 library.

* Implemented Model#pk_hash. Is it really necessary?

* Deprecated Model#pkey. Implemented better Model#pk method.

* Specs and docs for Model.one_to_one, Model.one_to_many macros.

=== 0.3.0.1 (2007-10-20)

* Changed Database#fetch to return a modified dataset.

=== 0.3 (2007-10-20)

* Added stock transforms to Dataset#transform. Refactored Model.serialize.

* Added Database#logger= method for setting the database logger object.

* Fixed Model.[] to act as shortcut to Model.find when a hash is given (#71).

* Added support for old and new decimal types in MySQL adapter, and updated MYSQL_TYPES with MySQL 5.0 constants (#72).

* Implemented Database#disconnect method for all adapters.

* Fixed small bug in ArrayKeys module.

* Implemented model caching by primary key.

* Separated Model.find and Model.[] functionality. Model.find takes a filter. Model.[] is strictly for finding by primary keys.

* Enhanced Dataset#first to accept a filter block. Model#find can also now accept a filter block.

* Changed Database#[] to act as shortcut to #fetch if a string is given.

* Renamed Database#each to #fetch. If no block is given, the method returns an enumerator.

* Changed Dataset#join methods to correctly literalize values in join conditions (#70).

* Fixed #filter with ranges to correctly literalize field names (#69).

* Implemented Database#each method for quickly retrieving records with arbitrary SQL (thanks Aman Gupta).

* Fixed bug in postgres adapter where a LiteralString would be literalized as a regular String.

* Fixed SQLite insert with subquery (#68).

* Reverted back to hashes as default mode. Added Sequel.use_array_tuples and Sequel.use_hash_tuples methods.

* Fixed problem with arrays with keys when using #delete.

* Implemented ArrayKeys as substitute for ArrayFields.

* Added Dataset#each_hash method.

* Rewrote SQLite::Database#transaction to use sqlite3-ruby library implementation of transactions.

* Fixed Model.destroy_all to work correctly in cases where no before_destroy hook is defined and an after_destroy hook is defined.

* Restored Model.has_hooks? implementation.

* Changed Database#<< to strip comments and whitespace only when an array is given.

* Changed Schema::Generator#primary_key to accept calls with the type argument omitted.

* Hooks can now be prepended or appended by choice.

* Changed Model.subset to define filter method on the underlying dataset instead of the model class.

* Fixed Dataset#transform to work with array fields.

* Added Dataset#to_csv method.

* PrettyTable can now extract column names from arrayfields.

* Converted ado, dbi, odbc adapters to use arrayfields instead of hashes.

* Fixed composite key support.

* Fixed Dataset#insert_sql, update_sql to support array fields.

* Converted sqlite, mysql, postgres adapters to use arrayfields instead of hashes.

* Extended Dataset#from to auto alias sub-queries.

* Extended Dataset#from to accept hash for aliasing tables.

* Added before_update, after_update hooks.

=== 0.2.1.1 (2007-10-07)

* Added Date literalization to sqlite adapter (#60).

* Changed Model.serialize to allow calling it after the class is defined (#59).

* Fixed after_create hooks to allow calling save inside the hook (#58).

* Fixed MySQL quoting of sql functions (#57).

* Implemented rollback! global method for cancelling transactions in progress.

* Fixed =~ operator in Sequelizer.

* Fixed ODBC::Dataset#fetch_rows (thanks Dusty).

* Renamed Model.recreate_table to create_table!. recreate_table is deprecated and will issue a warning (#56).

=== 0.2.1 (2007-09-24)

* Added default implementation of Model.primary_key_hash.

* Fixed Sequel::Model() to set dataset for inherited classes.

* Rewrote Model.serialize to use Dataset#transform.

* Implemented Dataset#transform.

* Added gem spec for Windows (without ParseTree dependency).

* Added support for dynamic strings in Sequelizer (#49).

* Query branch merged into trunk.

* Implemented self-changing methods.

* Add support for ternary operator to Sequelizer.

* Fixed sequelizer to evaluate expressions if they don't involve symbols or literal strings.

* Added protection against using #each, #delete, #insert, #update inside query blocks.

* Improved Model#method_missing to deal with invalid attributes.

* Implemented Dataset#query.

* Added Dataset#group_by as alias for Dataset#group.

* Added Dataset#order_by as alias for Dataset#order.

* More model refactoring. Added support for composite keys.

* Added Dataset#empty? method (#46).

* Fixed Symbol#to_field_name to support names with numbers and upper-case characters (#45).

* Added install_no_doc rake task.

* Partial refactoring of model code.

* Refactored dataset-model association and added Dataset#set_row_filter method.

* Added support for case-sensitive regexps to mysql adapter.

* Changed mysql adapter to support encoding option as well.

* Added charset/encoding option to postgres adapter.

* Implemented Model.serialize (thanks Aman Gupta.)

* Changed Model.create to INSERT DEFAULT VALUES instead of (id) VALUES (null) (brings back #41.)

* Fixed Model.new to work without arguments.

* Added Model.no_primary_key method to allow models without primary keys.

* Added Model#this method (#42 thanks Duane Johnson).

* Fixed Dataset#insert_sql to use DEFAULT VALUES clause if argument is an empty hash.

* Fixed Model.create to work correctly when no argument is passed (#41).

=== 0.2.0.2 (2007-09-07)

* Dataset#insert can now accept subqueries.

* Changed Migrator.apply to return the version.

* Changed Sequel::Model() to cache intermediate classes so descendant classes can be reopened (#39).

* Added :charset option to MySQL adapter (#40).

* Fixed Dataset#exclude to add parens around NOT expression (#38).

* Fixed use of sub-queries with all comparison operators in block filters (#38).

* Fixed arithmetic expressions in block filters to not be literalized.

* Changed Symbol#method_missing to return LiteralString.

* Changed PrettyTable to right-align numbers.

* Fixed Model.create_table (thanks Duane Johnson.)

=== 0.2.0.1 (2007-09-04)

* Improved support for invoking methods with inline procs inside block filters.

=== 0.2.0 (2007-09-02)

* Fixed Model.drop_table (thanks Duane Johnson.)

* Dataset#each can now return rows for arbitrary SQL by specifying :sql option.

* Added spec for postgres adapter.

* Fixed Model.method_missing to work with new SQL generation.

* Fixed #compare_expr to support regexps.

* Fixed postgres, mysql adapters to support regexps.

* More specs for block filters. Updated README.

* Added support for globals and $X macros in block filters.

* Fixed Sequelizer to not fail if ParseTree or Ruby2Ruby gems are missing.

* Renamed String#expr into String#lit (#expr should be deprecated in future versions).

* Renamed Sequel::ExpressionString into LiteralString.

* Fixed Symbol#[] to return an ExpressionString, so as not to be literalized.

* Renamed Dataset::Expressions to Dataset::Sequelizer.

* Renamed Expressions#format_re_expression to match_expr.

* Renamed Expressions#format_eq_expression to compare_expr.

* Added support for Regexp in MySQL adapter.

* Refactored Regexp expressions into a separate #format_re_expression method.

* Added support for arithmetic in proc filters.

* Added support for nested proc expressions, more specs.

* Added support for SQL function using symbols, e.g. :sum[:x].

* Fixed deadlock bug in ConnectionPool.

* Removed deprecated old expressions.rb.

* Rewrote Proc filter feature using ParseTree.

* Added support for additional functions on columns using Symbol#method_missing.

* Added support for supplying filter block to DB#[] method, to allow stuff like DB[:nodes] {:path =~ /^icex1/}.

=== 0.1.9.12 (2007-08-26)

* Added spec for PrettyTable.

* Added specs for Schema::Generator and Model (#36 thanks technoweenie).

* Fixed Sequel::Model.set_schema (#36 thanks technoweenie.)

* Added support for no options on Schema::Generator#foreign_key (#36 thanks technoweenie.)

* Implemented (restored?) Schema::Generator#primary_key_name (#36 thanks technoweenie.)

* Better spec code coverage.

=== 0.1.9.11 (2007-08-24)

* Changed Dataset#set_model to allow supplying additional arguments to the model's initialize method (#35). Thanks Sunny Hirai.

=== 0.1.9.10 (2007-08-22)

* Changed schema generation code to generate separate statements for CREATE TABLE and each CREATE INDEX (#34).

* Refactored Dataset::SQL#field_name for better support of different field quoting standards by specific adapters.

* Added #current_page_record_count for paginated datasets.

* Removed Database#literal and included Dataset::SQL instead.

* Sequel::Dataset:SQL#field_name can now take a hash (as well as #select and any method that uses #field_name) for aliasing column names.  E.g. DB[:test].select(:_qqa => 'Date').sql #=> 'SELECT _qqa AS Date FROM test'.

* Moved SingleThreadedPool to lib/sequel/connection_pool.rb.

* Changed SQLite::Dataset to return affected rows for #delete and #update (#33).

* ADO adapter: Added use of Enumerable for Recordset#Fields, playing it safe and moving to the first row before getting results, and changing the auto_increment constant to work for MSSQL.

=== 0.1.9.9 (2007-08-18)

* New ADO adapter by cdcarter (#31).

* Added automatic column aliasing to #avg, #sum, #min and #max (#30).

* Fixed broken Sequel::DBI::Dataset#fetch_rows (#29 thanks cdcarter.)

=== 0.1.9.8 (2007-08-15)

* Fixed DBI adapter.

=== 0.1.9.7 (2007-08-15)

* Added support for executing batch statements in sqlite adapter.

* Changed #current_page_record_range to return 0..0 for an invalid page.

* Fixed joining of aliased tables.

* Improved Symbol#to_field_name to prevent false positives.

* Implemented Dataset#multi_insert with :commit_every option.

* More docs for Dataset#set_model.

* Implemented automatic creation of convenience methods for each adapter (e.g. Sequel.sqlite etc.)

=== 0.1.9.6 (2007-08-13)

* Refactored schema definition code. Gets rid of famous primary_key problem as well as other issues (e.g. issue #22).

* Added #pagination_record_count, #page_range and #current_page_record_range for paginated datasets.

* Changed MySQL adapter to automatically reconnect (issue #26).

* Changed Sequel() to accept variable arity.

* Added :elements option to column definition, in order to support ENUM and SET types.

=== 0.1.9.5 (2007-08-12)

* Fixed migration docs.

* Removed dependency on PGconn in Schema class.

=== 0.1.9.4 (2007-08-11)

* Added Sequel.dbi convenience method for using DBI connection strings to open DBI databases.

=== 0.1.9.3 (2007-08-10)

* Added support for specifying field size in schema definitions (thanks Florian Aßmann.)

* Added migration code based on work by Florian Aßmann.

* Reintroduced metaid dependency. No need to keep a local copy of it.

=== 0.1.9.2 (2007-07-24)

* Removed metaid dependency. Re-factored requires in lib/sequel.rb.

=== 0.1.9.1 (2007-07-22)

* Improved robustness of MySQL::Dataset#field_name.

* Added Sequel.single_threaded= convenience method.

=== 0.1.9 (2007-07-21)

* Fixed #update_sql and #insert_sql to support field quoting by calling #field_name.

* Implemented automatic data type conversion in mysql adapter.

* Added support for boolean literals in mysql adapter.

* Added support for ORDER and LIMIT clauses in UPDATE statements in mysql adapter.

* Implemented correct field quoting (using back-ticks) in mysql adapter.

* Wrote basic MySQL spec.

* Fixd MySQL::Dataset to return correct data types with symbols as hash keys.

* Removed discunctional MySQL::Database#transaction.

* Added support for single threaded operation.

* Fixed bug in Dataset#format_eq_expression where Range objects would not be literalized correctly.

* Added parens around postgres LIKE expressions using regexps.

=== 0.1.8 (2007-07-10)

* Implemented Dataset#columns for retrieving the columns in the result set.

* Updated Model with changes to how model-associated datasets work.

* Beefed-up specs. Coverage is now at 95.0%.

* Added support for polymorphic datasets.

* The adapter dataset interface was simplified and standardized. Only four methods need be overriden: fetch_rows, update, insert and delete.

* The Dataset class was refactored. The bulk of the dataset code was moved into separate modules.

* Renamed Dataset#hash_column to Dataset#to_hash.

* Added some common pragmas to sqlite adapter.

* Added Postgres::Dataset#analyze for EXPLAIN ANALYZE queries.

* Fixed broken Postgres::Dataset#explain.

=== 0.1.7

* Removed db.synchronize wrapping calls in sqlite adapter.

* Implemented Model.join method to restrict returned columns to the model table (thanks Pedro Gutierrez).

* Implemented Dataset#paginate method.

* Fixed after_destroy hook.

* Improved Dataset#first and #last to accept a filter hash.

* Added Dataset#[]= method.

* Added Sequel() convenience method.

* Fixed Dataset#first to include a LIMIT clause for a single record.

* Small fix to Postgres driver to return a primary_key value for the inserted record if it is specified in the insertion values (thanks Florian Aßmann and Pedro Gutierrez).

* Fixed Symbol#DESC to support qualified notation (thanks Pedro Gutierrez).

=== 0.1.6

* Fixed Model#method_missing to raise for an invalid attribute.

* Fixed PrettyTable to print model objects (thanks snok.)

* Fixed ODBC timestamp conversion to return DateTime rather than Time object (thanks snok.)

* Fixed Model.method_missing (thanks snok.)

* Model.method_missing now creates stubs for calling Model.dataset methods. Methods like Model.each etc are removed.

* Changed default join type to INNER JOIN (thanks snok.)

* Added support for literal expressions, e.g. DB[:items].filter(:col1 => 'col2 - 10'.expr).

* Added Dataset#and.

* SQLite adapter opens a memory DB if no database is specified, e.g. Sequel.open 'sqlite:/'.

* Added Dataset#or, pretty nifty.

=== 0.1.5

* Fixed Dataset#join to support multiple joins. Added #left_outer_join, #right_outer_join, #full_outer_join, #inner_join methods.

=== 0.1.4

* Added String#split_sql.

* Implemented Array#to_sql and String#to_sql. Database#to_sql can now take an array of strings and convert into an SQL string. Comments and excessive white-space are removed.

* Improved Schema generator to support data types as method names:
  DB.create_table :test do
    integer :abc
    text :def
    ...
  end

* Implemented ODBC adapter.

=== 0.1.3

* Implemented DBI adapter.

* Refactored database connection code. Now handled through Database#connect.

=== 0.1.2

* The first opened database is automatically assigned to to Model.db.

* Removed SequelConnectionError. Exception class errors are converted to RuntimeError.

* Added support for UNION, INTERSECT and EXCEPT set operations.

* Fixed Dataset#single_record to return nil if no record is found.

* Updated specs to conform to RSpec 1.0.

* Added Model#find_or_create method.

* Fixed MySQL::Dataset#query_single (thanks Dries Harnie.)

* Added Model.subset method. Fixed Model.filter and Model.exclude to accept blocks.

* Added Database#uri method.

* Refactored and removed deprecated code in postgres adapter.

===0.1.1

* More documentation for Dataset.

* Added Dataset#size as alias to Dataset#count.

* Changed Database#<< to call execute (instead of being an alias). Thus it will work for descendants as well.

* Fixed Sequel.open to accept variable arity.

* Refactored Model#refresh, Model.create. Removed Model#reload.

* Refactored Model hooks.

* Cleaned up Dataset API.

=== 0.1.0

* Changed Database#create_table to only accept a block. Nobody's gonna use the other way.

* Removed Dataset#[]= method. Too confusing and not really useful.

* Fixed ConnectionPool#hold to wrap exceptions only once.

* Dataset#where_list Renamed Dataset#expression_list.

* Added support for qualified fields in Proc expressions (e.g. filter {items.id == 1}.)

* Added like? and in? Proc expression operators.

* Added require 'date' in dataset.rb. Is this a 1.8.5 thing?

* Refactored Dataset to use literal strings instead of format strings (slight performance improvement and better readability.)

* Added support for literalizing Date objects.

* Refactored literalization of Time objects.

=== 0.0.20

* Refactored Dataset where clause construction to use expressions.

* Implemented Proc expressions (adapted from a great idea by Sam Smoot.)

* Fixed Model#map.

* Documentation for ConnectionPool.

* Specs for Database.

=== 0.0.19

* More specs for Dataset.

* Fixed Dataset#invert_order to work correctly with strings.

* Fixed Model#== to check equality of values.

* Added Model#exclude and Model#order.

* Fixed Dataset#order and Dataset#group to behave correctly when supplied with qualified field name symbols.

* Removed Database#literal. Shouldn't have been there.

* Added SQLite::Dataset#explain. Returns an array of opcode hashes.

* Specs for ConnectionPool.

=== 0.0.18

* Implemented SequelError and SequelConnectionError classes. ConnectionPool#hold now catches any connection errors and reraises them SequelConnectionError.

* Removed duplication in Database#[].

* :from and :select options are now always arrays (patch by Alex Bradbury.)

* Fixed Dataset#exclude to work correctly (patch and specs by Alex Bradbury.)

=== 0.0.17

* Fixed Postgres::Database#tables to return table names as symbols (caused problem when using Database#table_exists?).

* Fixed Dataset#from to have variable arity, like Dataset#select and Dataset#where (patch by Alex Bradbury.)

* Added support for GROUP BY and HAVING clauses (patches by Alex Bradbury.) Refactored Dataset#filter.

* More specs.

* Refactored Dataset#where for better composability.

* Added Dataset#[]= method.

* Added support for DISTINCT and OFFSET clauses (patches by Alex Bradbury.) Dataset#limit now accepts ranges. Added Dataset#uniq and distinct methods.

=== 0.0.16

* More documentation.

* Added support for subqueries in Dataset#literal.

* Added support for Model.all_by_XXX methods through Model.method_missing.

* Added basic SQL logging to Database.

* Added Enumerable#send_each convenience method.

* Changed Dataset#destroy to return the number of deleted records.

=== 0.0.15

* Improved Dataset#insert_sql to allow arrays as well as hashes.

* Database#drop_table now accepts a list of table names.

* Added Model#id to to return the id column.

=== 0.0.14

* Fixed Model's attribute accessors (hopefully for the last time).

* Changed Model.db and Model.db= to allow different databases for different model classes.

* Fixed bug in aggregate methods (max, min, etc) for datasets using record classes.

=== 0.0.13

* Fixed Model#method_missing to do both find, filter and attribute accessors. duh.

* Fixed bug in Dataset#literal when quoting arrays of strings (thanks Douglas Koszerek.)

=== 0.0.12

* Model#save now correctly performs an INSERT for new objects.

* Added Model#reload for reloading an object from the database.

* Added Dataset#naked method for getting a version of a dataset that fetches records as hashes.

* Implemented attribute accessors for column values ala ActiveRecord models.

* Fixed filtering using nil values (e.g. dataset.filter(:parent_id => nil)).

=== 0.0.11

* Renamed Model.schema to Model.set_schema and Model.get_schema to Model.schema.

* Improved Model class to allow descendants of model clases (thanks Pedro Gutierrez.)

* Removed require 'postgres' in schema.rb (thanks Douglas Koszerek.)

=== 0.0.10

* Added some examples.

* Added Dataset#print method for pretty-printing tables.

=== 0.0.9

* Fixed Postgres::Database#tables and #locks methods.

* Added PGconn#last_insert_id method that should support all 7.x and 8.x versions of Postgresql.

* Added Dataset#exists method for EXISTS where clauses.

* Changed behavior of Dataset#literal to regard symbols as field names.

* Refactored and DRY'd Dataset#literal and overrides therof. Added support for subqueries in where clause.

=== 0.0.8

* Fixed Dataset#reverse_order to provide chainability. This method can be called without arguments to invert the current order or with arguments to provide a descending order.

* Fixed literal representation of literals in SQLite adapter (thanks Christian Neukirchen!) 

* Refactored insert code in Postgres adapter (in preparation for fetching the last insert id for pre-8.1 versions).

=== 0.0.7

* Fixed bug in Model.schema, duh!

=== 0.0.6

* Added Dataset#sql as alias to Dataset#select_sql.

* Dataset#where and Dataset#exclude can now be used for refining dataset conditions, enabling stuff like posts.where(:title => 'abcdef').exclude(:user_id => 3).

* Implemented Dataset#exclude method.

* Added Sequel::Schema#auto_primary_key method for setting an automatic primary key to be added to every table definition. Changed the schema generator to not define a primary key by default.

* Changed Sequel::Database#table_exists? to rely on the tables method if it is available.

* Implemented SQLite::Database#tables.

=== 0.0.5

* Added Dataset#[] method. Refactored Model#find and Model#[].

* Renamed Pool#conn_maker to Pool#connection_proc.

* Added automatic require 'sequel' to all adapters for convenience.

=== 0.0.4

* Added preliminary MySQL support.

* Code cleanup.

=== 0.0.3

* Add Dataset#sum method.

* Added support for exclusive ranges (thanks Christian Neukirchen.)

* Added sequel console for quick'n'dirty access to databases.

* Fixed small bug in Dataset#qualified_field_name for better join support.

=== 0.0.2

* Added Sequel.open as alias to Sequel.connect.

* Refactored Dataset#where_equal_condition into Dataset#where_condition, allowing arrays and ranges, e.g. posts.filter(:stamp => (3.days.ago)..(1.day.ago)), or posts.filter(:category => ['ruby', 'postgres', 'linux']).

* Added Model#[]= method for changing column values and Model#save 
method for saving them.

* Added Dataset#destroy for deleting each record individually as support for models. Renamed Model#delete to Model#destroy (and Model#destroy_all) ala ActiveRecord.

* Refactored Dataset#first and Dataset#last code. These methods can now accept the number of records to fetch.

=== 0.0.1

* More documentation for Dataset.

* Renamed Database#query to Database#dataset.

* Added Dataset#insert_multiple for inserting multiple records.

* Added Dataset#<< as shorthand for inserting records.

* Added Database#<< method for executing arbitrary SQL.

* Imported Sequel code.


== Sequel::Model CHANGELOG 0.1 - 0.5.0.2

=== 0.5.0.2 (2008-03-12)

* More fixes for Model.associate to accept strings and symbols as class references.

=== 0.5.0.1 (2008-03-09)

* Fixed Model.associate to accept class and class name in :class option.

=== 0.5 (2008-03-08)

* Merged new associations branch into trunk.

* Rewrote RDoc for associations.

* Added has_and_belongs_to_many alias for many_to_many.

* Added support for optional dataset block.

* Added :order option to order association datasets.

* Added :cache option to return and cache array of objects for association.

* Changed one_to_many, many_to_many associations to return dataset by default.

* Added has_many, belongs_to aliases.

* Refactored associations code.

* Added deprecations for old-style relations.

* Completed specs for new associations code.

* New associations code by Jeremy Evans (replaces relations code.)

=== 0.4.2 (2008-02-29)

* Fixed one_to_many implicit key to work correctly for namespaced classes (#167).

* Fixed Model.db= to affect the underlying dataset (#183).

* Fixed Model.implicit_table_name to disregard namespaces.

=== 0.4.1 (2008-02-10)

* Implemented Model#inspect (#151).

* Changed Model#method_missing to short-circuit and bypass checking #columns if the values hash already contains the relevant column (#150).

* Updated to reflect changes in sequel_core (Dataset#clone_merge renamed to Dataset#clone).

=== 0.4 (2008-02-05)

* Fixed Model#set to work with string keys (#143).

* Fixed Model.create to correctly initialize instances marked as new (#135).

* Fixed Model#initialize to convert string keys into symbol keys. This also fixes problem with validating objects initialized with string keys (#136).

=== 0.3.3 (2008-01-25)

* Finalized support for virtual attributes.

=== 0.3.2.1 (2008-01-24)

* Fixed Model.dataset to correctly set the dataset if using implicit naming or inheriting the superclass dataset (thanks celldee).

=== 0.3.2 (2008-01-24)

* Added Model#update_with_params method with support for virtual attributes and auto-filtering of unrelated parameters, and changed Model.create_with_params to support virtual attributes (#128).

* Cleaned up gem spec (#132).

* Removed validations code. Now relying on validations in assistance gem.

=== 0.3.1 (2008-01-21)

* Changed Model.dataset to use inflector to pluralize the class name into the table name. Works in similar fashion to table names in AR or DM.

=== 0.3 (2008-01-18)

* Implemented Validatable::Errors class.

* Added Model#reload as alias to Model#refresh.

* Changed Model.create to accept a block (#126).

* Rewrote validations.

* Fixed Model#initialize to accept nil values (#115).

=== 0.2 (2008-01-02)

* Removed deprecated Model.recreate_table method.

* Removed deprecated :class and :on options from one_to_many macro.

* Removed deprecated :class option from one_to_one macro.

* Removed deprecated Model#pkey method.

* Changed dependency to sequel_core.

* Removed examples from sequel core.

* Additional specs. We're now at 100% coverage.

* Refactored hooks code. Hooks are now inheritable, and can be defined by supplying a block or a method name, or by overriding the hook instance method. Hook chains can now be broken by returning false (#111, #112).

=== 0.1 (2007-12-30)

* Moved model code from sequel into separate model sub-project.